OpenAI has code red again. And they're urgently changing strategy.
At an all-hands meeting, head of applications Fidji Simo announced a course change for OpenAI. The emphasis is shifting to enterprise users and developers — the race for the consumer market appears to be slowing: "OpenAI can't be distracted by side quests, we need hard focus on business and productivity." She called Anthropic and Claude's success the trigger (wake-up call).
What consumer projects are at risk:
- the Sora-based social network (TikTok for AI video)
- AI earbuds (that Altman and Jony Ive project)
- the ChatGPT Atlas browser
What they're strengthening:
- enterprise solutions
- API and developer tools
- agent applications (GPT-5.4 is already optimized for this, natively working across multiple apps on one machine)
In short, OpenAI got scared that Anthropic would take the entire enterprise market and decided not to spread thin. They put consumer experiments on pause.
UPDATE: Decided to add: today OpenAI signed a deal with AWS to deliver AI services to US government agencies. Last week they acquired Promptfoo — a startup focused on AI agent safety. And they're negotiating a $10B joint venture with TPG, Bain Capital, and Brookfield to distribute enterprise AI through their portfolios. Strategy in action.